skills/claude-skills-open/skills/sales/call-prep/SKILL.md
--- name: call-prep description: Call preparation: research, CRM, talking points, PDF --- # Call Prep > Preparation for a call with a client/lead: research, CRM update, conversation plan, PDF ## When to use - "prepare for a call with X" - "call prep for Y" - "gather info before a meeting" - "conversation plan with a client" - There is a scheduled call in the calendar or task ## Dependencies - CRM data: `query-leads` - Web search: WebSearch, WebFetch - PDF: `weasyprint` (Python) ## Paths |
npx skillsauth add aaaaqwq/agi-super-skills skills/claude-skills-open/skills/sales/call-prepInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Preparation for a call with a client/lead: research, CRM update, conversation plan, PDF
query-leadsweasyprint (Python)| What | Path |
|------|------|
| CRM Companies | $CRM_PATH/contacts/companies.csv |
| CRM People | $CRM_PATH/contacts/people.csv |
| CRM Leads | $CRM_PATH/relationships/leads.csv |
| CRM Activities | $CRM_PATH/activities.csv |
| PM Tasks | $PM_PATH/pm_tasks_master.csv |
| Output PDF | $PROJECT_ROOT/docs/{slug}-call-prep.pdf |
Read everything from CRM about this person/company:
1. companies.csv -- company record
2. people.csv -- person record + notes
3. leads.csv -- lead stage, priority, next_action, notes
4. activities.csv -- communication history (emails, calls, messages)
5. pm_tasks_master.csv -- related tasks
Important: gather ALL interaction history -- not just the latest entry.
Run in parallel:
1. WebSearch: "{name} {company}" -- general info
2. WebSearch: "{name} linkedin founder" -- career, track record
3. WebFetch: company website -- products, positioning, pricing
4. WebSearch: "{company} 2025 2026" -- latest news
5. WebFetch: LinkedIn profile (if URL exists in CRM)
What to look for:
Based on research, update:
companies.csv -- description, industry, sizepeople.csv -- role, notes from researchleads.csv -- notesUse skill update-lead.
If a client workspace exists in Drive (Clients/{CompanyName}/), check:
DM="$GOOGLE_TOOLS_PATH/.venv/bin/python3 $GOOGLE_TOOLS_PATH/drive_manager.py"
# Search for client folder
$DM search "CompanyName" --folder <YOUR_CLIENTS_FOLDER_ID>
# List docs in client folder
$DM list CLIENT_FOLDER_ID
# For each shared doc — check if client opened/edited it
$DM info DOC_ID
# → Look at lastModifiedBy: if it's the client, they filled it in
# → Look at modifiedTime: when was it last touched
If questionnaire exists and client filled it in: read their answers and incorporate into conversation plan.
If questionnaire exists but client didn't fill it: mention on the call, go through questions verbally.
If no workspace exists: consider creating one with client-workspace skill.
See skill: client-workspace
Standard discovery call structure:
Phase 1: Small talk + context (~3 min)
- How you got in touch
- What you know about them (but not everything -- let them tell)
Phase 2: Business discovery (~10 min)
- What does the company do?
- What products/services?
- Who are the customers?
- What stage? (pre-launch, growth, scaling)
- How many people on the team?
Phase 3: Pain point discovery (~10 min)
- What specifically hurts? What problem do they want to solve?
- What have they already tried?
- What didn't work and why?
- What is the budget/expectations?
- What are the deadlines?
Phase 4: Show relevance (~5 min)
- Specific example of how we solved a similar task
- Don't sell -- show that you understand the problem
- Adapt to the level of the conversation partner
Phase 5: Propose a format (~5 min)
- Option A: Audit (5-10h, understand scope)
- Option B: Pilot (fixed task, 2 weeks)
- Option C: Partnership (after pilot)
- DO NOT name a price without scope
Phase 6: Next steps (~2 min)
- Specific next step
- Deadline
- What is needed from them
Typical risks:
| Risk | How to respond | |------|---------------| | Wants to "look" for free | An audit is work. Minimum paid. | | Scope too large | Narrow down to one project/channel | | Wants equity deal right away | Paid pilot first | | Already has a solution, comparing | Ask who they're comparing with | | Not the decision maker | Ask who makes the decision | | No budget | Propose a minimal pilot |
Create HTML with all info, convert to PDF via weasyprint:
import weasyprint
html = "..." # structured HTML with sections 1-5
weasyprint.HTML(string=html).write_pdf('/path/to/output.pdf')
PDF structure:
Open PDF:
open /path/to/output.pdf
User: prepare for a call with Alisa from shftd.ai
Claude:
1. Reads CRM → comp-shftd, p-shftd-001, lead-shftd-001
2. WebSearch "Alisa Chumachenko shftd.ai" → Game Insight, GOSU.ai, Forbes
3. WebFetch shftd.ai → pre-launch, venture studio
4. Updates CRM with research
5. Builds plan: discovery call, 6 phases, specific questions
6. Generates PDF → docs/alisa-shftd-call-prep.pdf
User: prepare for a follow-up with Client F Tech
Claude:
1. Reads CRM → comp-clientf, lead, activities (history)
2. Checks what was discussed earlier
3. Builds plan: review progress, discuss blockers, next steps
4. Generates PDF
pip install weasyprint)query-leads -- reading CRM dataupdate-lead -- updating CRM with researchdaily-briefing -- may contain info about scheduled callsgit-workflow -- commit CRM changes after updatetesting
AI驱动的智能浏览器自动化工具。使用LLM理解页面并自动执行任务,比传统Playwright更智能、更省token。适用于复杂交互、动态页面、需要智能决策的浏览器操作。Chrome浏览器优先。
tools
网页登录态管理。使用 fast-browser-use (fbu) 管理各平台登录状态,定期检查可用性,新平台授权时自动保存 profile。
development
Monitor and report on API provider quotas, balances, and usage. Query official providers (Moonshot, DeepSeek, xAI, Google AI Studio) and relay/proxy providers (Xingjiabiapi, Aixn, WoW) via their billing APIs. Also checks subscription services (Brave Search, OpenRouter). Generates quota reports. Triggers on "查额度", "API余额", "quota check", "billing report", "api balance", "供应商额度", "中转站余额", "费用报告", "check balance", "how much credit".
development
# A股基金监控 Skill A股基金净值监控,支持实时估值和盘后净值,自动判断交易日/节假日。 ## 用法 ### 快速监控(命令行) ```bash # 默认配置,输出到控制台 bash ~/clawd/skills/a-fund-monitor/scripts/monitor.sh # 推送到群(使用--push参数) bash ~/clawd/skills/a-fund-monitor/scripts/monitor.sh --push # 监控指定基金 bash ~/clawd/skills/a-fund-monitor/scripts/monitor.sh --codes "000979 002943" ``` ### Agent调用 ``` 执行A股基金监控任务。 1. 读取配置文件: ~/clawd/skills/a-fund-monitor/config.json 2. 获取实时净值数据 3. 非交易日自动切换为简短报告 配置文件格式: { "funds": [ {"code": "000979", "name": "景顺长城沪港深精选股票